MicroAI Factory Offers On Device ML-Based Solution for OEE

In March 2022, Dallas-based MicroAI announced the general availability of its MicroAI Factory solution. MicroAI Factory ingests machine and operational data to improve Overall Equipment Effectiveness (OEE), real-time cycle time analysis, and predictive maintenance. MicroAI Factory is deployed in commercial servers and designed to be plug-and-play. These servers can be connected across multiple facilities or production sites to share insights and further enhance the accuracy of the Machine Learning (ML) solution.

Data collected from Programmable Logic Controllers (PLCs) and machine sensors, including pressure, vibration, temperature, and audio, are analyzed using an ML inference engine deployed locally within an industrial computing appliance at a manufacturing or industrial site. The ML engine will generate cycle time analysis and behavioral models that provide real-time performance, productivity, and uptime analysis

Using Machine Learning to provide insight into OEE is becoming more common. By leveraging data collected from various production machinery, manufacturers can make sense of the current state of their operation, significantly reducing downtime, poor product quality, and cost inefficiency.

Founded in 2018, MicroAI was formerly known as One Tech. The company develops a series of ML-based software and solutions that can understand the behavior of devices, equipment, and machines. MicroAI’s ML solution trains and runs locally without sending any data to the cloud, making them secure from any tampering. This is crucial for manufacturers that prefer to have data stored locally. In addition, ML training can be performed without the need to label incoming sensor data. Most importantly, MicroAI’s ML software does not require high computing resources and is designed to be hardware agnostic so that it can be deployed as embedded software in any industrial device powered by Microcontrollers (MCUs). This means legacy equipment that does not have a dedicated ML chipset can also be embedded with MicroAI software. Such implementation brings accurate, ultra-low latency and real-time prediction outputs while protecting privacy and cybersecurity.

As the name implies, MicroAI Factory is designed for an entire manufacturing or production site. For smaller scale requirements, MicroAI also offers MicroAI AtomML and MicroAI Network. AtomML is an ML algorithm that can run on resource-constrained environments for real-time analysis of time-series data for a single industrial device, while Network is for a gateway that connects to multiple devices. In addition, MicroAI also offers MicroAI Security which protects and monitors the device’s behavior from a cybersecurity standpoint. Key features include event monitoring, anomaly detection, and vulnerability mitigation.

The predictive maintenance and OEE market is a fast-growing one. According to ABI Research’s Digital Factory Data (MD-IICT-107), the global Internet of Things (IoT) analytics and data management services revenue are estimated to be US$7.7 billion in 2022. This is expected to grow to US$19.7 billion by 2026. Automotive, electronics, machinery, and food and beverage are leading the digital transformation in the manufacturing sector, but other verticals are set to catch up soon.

Sensing the opportunities, different players are introducing various products targeting this domain. This includes industrial equipment vendors such as Rockwell Automation, Siemens, and Schneider Electric, public cloud vendors such as Amazon Web Services (AWS), Google, and Microsoft, and industrial IoT platform vendors such as C3, FogHorn, PTC, and Falkonry. ABI Research has profiled some of these vendors in our past report on industrial AI platforms and service providers (CA-1269).

In comparison, MicroAI has offered key propositions that are important for end users. First, most manufacturers prefer to keep their operational data processed and stored on-premises. MicroAI’s solutions are designed to be fully localized and do not rely on cloud connectivity. Second, MicroAI has an in-depth understanding of the manufacturing sector and industrial machinery and equipment through close collaboration with Renesas, Honeywell, and HP. Finally, MicroAI can be scaled easily through a diverse range of solutions and works with all major Product Lifecycle Management (PLM) and Enterprise Resource Management (ERP) software and public cloud services. No doubt, more strategic collaboration and partnerships would make MicroAI very appealing to the end users looking for a low powered yet highly accurate on-device and on-premises ML solution to maximize the useful life of their legacy and existing industrial infrastructure. Considering these strengths, ABI Research expects MicroAI to be a key player in the OEE space moving forward.